In plain terms: Ship designs are among a yard’s most sensitive and valuable
assets, and some are defense-related. Forge is built so that sensitive data is
processed on your own infrastructure and never leaves the building, and so that no
yard can ever see another yard’s work.
Why this is the hardest constraint
An AI is only as good as the data it learns from, which means feeding it a yard’s past designs and contracts. But that data is also the yard’s most valuable and sensitive asset. Sending proprietary hull forms and regulatory details to a shared cloud model is a risk most yards cannot accept. So Forge treats data sensitivity as a versioned policy, not a runtime preference. It is enforced by the system, not left to good intentions.Four sensitivity classes
Every source and every AI request carries one of four sensitivity classes, and the class decides where it is allowed to be processed:| Class | What it covers | Where it can go |
|---|---|---|
| Public | Reference rule text, public datasets, published papers | Any model provider |
| Internal | Forge’s own schemas, prompts, evaluations | Any provider |
| Confidential | Tenant-provided non-defense data, commercial yard records | Local (your own infrastructure) by default; external only with explicit redaction and a fidelity check |
| Restricted | USCG, ABS, defense-adjacent, IP-worth implementation specifics | Local only. Never external. All outside providers blocked. |
“Zero bytes of restricted data crossed the boundary” is something you can check, not just something you hope.